Tec Nikan
فارسی
Talk to us
All news

In-Line X-Ray Inspection Arrives at Conveyor Speed

A new arrayed system inspects parts at up to 760 mm per second, aimed at 100% inspection rather than sampling, with analysis running in the cloud.

metrologyquality inspectionX-raymanufacturingindustrial data

Lumafield announced Mars on 22 July 2026, an in-line arrayed X-ray inspection system it says is available now and purpose-built to drop into existing production lines without custom fixturing or stopping production flow. It inspects at speeds up to 760 mm per second — 2.5 feet — and is aimed at 100% part inspection rather than statistical sampling.

The system is offered in 2D and 2.5D arrayed configurations, with the multilinear 2.5D architecture capturing depth information in a single pass for richer defect visibility. Hardware streams directly into the company's cloud-based analysis platform, which applies AI to the X-ray data. Co-founder and chief technology officer Andreas Bastian framed the differentiator as software, noting that in-line X-ray has existed for decades but was limited by the intelligence of the software reading it. Named target applications are electronics, automotive components, batteries, consumer packaged goods, and identifying counterfeit or defective returns.

Two consequences follow for anyone integrating this rather than selling it.

The first is data volume and timing. A high-bandwidth imaging source on a moving conveyor has to be time-synchronised with PLC and MES part-tracking data, or the images cannot be attributed to the parts they came from — which makes them useless for anything except a live pass/fail decision. That synchronisation is the unglamorous prerequisite and it is where these installations most often fall short.

The second is that the analysis lives in the cloud, which turns a quality decision into a network dependency. Egress bandwidth, latency and data residency all become production concerns rather than IT ones, and somebody has to decide what the line does when the link is down.

The shift from sampled to continuous inspection also changes the shape of downstream quality databases. A system designed around a few measured parts per batch behaves differently when every part has a record, and the retention question — how long to keep images versus derived measurements — arrives immediately.

Source: Lumafield

Want to work with us?

Tell us what you're building and we'll help you scope the first deployment.